A Visualization System for Performance Analysis of Image Classification Models
Chanhee Park, Hyojin Kim, Kyungwon Lee · Electronic Imaging · 2020
Building machine learning models for the task of image classification involves various tasks such as model selection, layer design, and hyperparameter tuning to improve model performance.However, regarding deep learning models, insufficient model interpretability renders it infeasible to understand how they make predictions.To facilitate model interpretation, performance analysis at the class and instance levels with model visualization is essential.We herein present an interactive visual analytics system to provide a wide range of performance evaluation methods of different machine learning models for image classification problems.The proposed system aims to overcome challenges by providing visual performance analysis at different levels and visualizing misclassification instances.The system which comprises five views -ranking, projection, matrix, and instance list views, enables the comparison and analysis different models through user interaction.We describe several use cases with multiple machine learning models on MNIST dataset to demonstrate effectiveness of the proposed system.Our demo app is available at https://chanhee13p.github.io/VisMlic/.